Solution to the Eigenproblem by a Norm Reducing Jacobi Type Method
Let A be an n×n real matrix. A matrix T=T 1 T 2…T i … (or T -1) is constructed as a product of a sequence of two dimensional transformations T i . From A′,= T -1 AT the eigenvalues may be read off and from T (or T -1) the right (or left) vectors. Each T i is of the form RS where R is a rotation and S a shear or complex rotation.
Unable to display preview. Download preview PDF.
- 2.Eberlein, P. J. Algorithms for the scaling of inner products with applications to linear least squares and the eigenproblem. Submitted to the SI AM J. Numer. Anal.Google Scholar